146 little red dots point to black holes wrapped in gas

In James Webb Space Telescope images, some sources in the young universe fit into only a few pixels and appear intensely red. Their appearance inspired competing explanations: exceptionally dense galaxies, active nuclei hidden by dust, or growing black holes inside thick gas? A study finalized September 3 in Monthly Notices of the Royal Astronomical Society assembled 146 of these “little red dots.” That sample and a review published the next day in Nature Astronomy converge on the third interpretation, although the sources’ internal structure is still inferred from spectra rather than photographed directly.

Anna de Graaff and colleagues began with roughly 17,000 high-quality spectra from the public archive of NIRSpec, Webb’s spectrograph. Their first filter sought a V-shaped continuum: relatively blue ultraviolet light followed by a red rise in the optical. The criterion returned 333 spectra. The team then required a compact source in 4.44-micrometer images from the NIRCam camera, removed problematic observations and consolidated duplicates. That left 146 unique objects from redshift 2.0 to 9.3, a range spanning several stages of the early universe.

The most direct signature came from hydrogen lines. Fifty-eight objects had medium- or high-resolution spectra; in 53 of them, the observed range covered H-alpha. A formal comparison of models with and without a broad component detected broad Balmer lines in 52 of the 53, or 98%. Fast-moving gas near a compact source commonly causes such broadening and is a sign of black-hole accretion. The 98% rate measures the purity of the selection among objects with suitable data; it does not mean the survey found 98% of all little red dots that exist.

The continuum shape supplied an independent clue. For objects with enough wavelength coverage, empirical modified-blackbody models described the optical and near-infrared light well. Fitted temperatures ranged from about 2,000 to 7,000 kelvins and clustered near 5,000 kelvins. H-alpha luminosity also increased in a tight, roughly linear relation with the optical continuum, while properties of other lines varied in coordinated ways. That chain indicates that the broad lines and much of the continuum draw energy from the same central source.

The proposed interpretation is an accreting massive black hole almost completely covered by a column of dense, mostly neutral gas. The gas would absorb and reprocess radiation from the disk, producing an apparent thermal spectrum near 5,000 kelvins and hiding some X-rays. Kohei Inayoshi and Luis Ho’s review regards black-hole masses of about one million to ten million Suns, inferred from the broad lines, as plausible. It also finds that stars are unlikely to dominate the entire output, although star formation in the host galaxy may contribute especially in ultraviolet light and to the ionized-oxygen [O III] line. “Black hole star” names the model; it is not a new class of star.

The sample was built to favor typical, pure sources rather than measure a complete population. Programs in the archive have different depths and target criteria, so fainter objects or those lacking the V shape may have been missed. Only 40% of the 146 objects had higher-resolution spectra, and the most reliable thermal fit required coverage on both sides of the peak, restricting part of the analysis to redshifts below 4.5. Broad lines may also receive a contribution from scattering within the gas, making mass estimates dependent on the model.

The 146 sources organize a problem that previously rested on scattered cases. Repeated broad lines, the H-alpha–continuum relation and the narrow temperature range favor an accretion engine wrapped in gas; the diversity of [O III] preserves a role for host galaxies. If that reading is right, little red dots record a brief phase in which black-hole seeds grew rapidly while remaining nearly sealed inside their fuel. Variability, ionizing emission, descendants at later times and nearby analogs are the tests that can turn the leading model into a complete physical history.